| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Experimental Studies |
1 From the Department of Radiology, Seoul National University College of Medicine and the Institute of Radiation Medicine, SNUMRC, Clinical Research Institute, 28 Yongon-dong, Chongno-gu, Seoul 110-744, Korea (J.M.G., J.G.I., H.J.L., Y.J.L., J.W.K., J.H.K.); Department of Radiology, Eulgi University School of Medicine, Seoul, Korea (M.J.C.); Department of Radiology, University of Ulsan, Asan Medical Center, Seoul, Korea (J.B.S.); and Department of Radiology, National Cancer Center, Goyang-si, Korea (H.Y.K.). Received August 29, 2001; revision requested October 12; revision received November 20; accepted January 22, 2002. Address correspondence to J.G.I. (e-mail: imjg@radcom.snu.ac.kr).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: To test the diagnostic performance of these two systems, four types of simulated lesions (nodules, micronodules, lines, and reticular opacities) were superimposed over an anthropomorphic chest phantom. Digital chest radiographs were acquired with amorphous silicon flat-panel–detector (3-K [K = 1,000] matrix, 12 bits) and storage-phosphor radiography (4-K matrix, 10 bits). Six board-certified radiologists evaluated soft-copy images on a high-resolution video monitor (2,048 x 2,560 x 8 bits). A total of 14,400 observations were analyzed in terms of receiver operating characteristics.
RESULTS: Average performance in terms of nodule detection was significantly better (P < .05) with the flat-panel–detector system than with the storage-phosphor system. For micronodules, lines, and reticular opacities, no significant detection differences in averaged performance were found between the two detector systems.
CONCLUSION: In the evaluation of soft-copy images, the amorphous silicon detector system appears to be superior to the storage-phosphor system for the detection of pulmonary nodules.
© RSNA, 2002
Index terms: Phantoms, 60.1215 • Radiography, comparative studies, 60.1215 • Radiography, digital, 60.1215 • Radiography, flat panel, 60.1215 • Thorax, radiography, 60.1215
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
For almost 20 years, storage-phosphor radiography has been providing digital images for projection radiography. The technique provides images that have the same quality as those produced with conventional screen-film systems (1,2). However, a new generation of digital flat-panel detectors has been developed, and various types of thin-film transistor detectors have been studied (3–5). All of these detectors are based on amorphous silicon thin-film transistor technology, but each is combined with different types of converter arrays, which convert the x-ray beams to electric charges either directly or indirectly. Physical indicators such as detective quantum efficiency can indicate overall system performance; however, no physical measurement correlates perfectly with perceived diagnostic quality.
The most important clinical criterion for the use of PACS is its enabling achievement of acceptable accuracy when radiologic images are interpreted at a soft-copy workstation. However, comparative studies (6) of soft-copy images obtained with various digital systems are rare. Moreover, the ability of radiologists to detect lesions determines whether a new imaging technique is superior to existing ones. Therefore, the findings of observer-based studies (5) with separate presentations of digital imaging systems are probably the most conclusive.
The aim of this study was to compare observer performance by using soft-copy images produced with an amorphous silicon flat-panel–detector system and with a storage-phosphor system for the detection of simulated chest lesions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
Digital Detector System Description
Posteroanterior chest radiographs were obtained by using two digital detector systems. Storage-phosphor images were obtained (model FCR-5501; Fuji, Tokyo, Japan) by using 35 x 43-cm imaging plates (model ST-55; Fuji), a 3,520 x 4,280 (4K [K = 1,000]) x 10-bit matrix, and a 0.1-mm pixel size. Flat-panel–detector system images were obtained (model JDC-9001; Phoenix Vision, Seoul, Korea) by using a 42.6 x 43.2-cm solid-state detector (Pixium 4600; Trixell, Moirans, France). The detector panel is fabricated on a monolithic glass substrate. An amorphous silicon thin-film transistor array is layered on the glass and is itself overlaid with a structured cesium iodide scintillator. X-ray beams are converted to visible light by the scintillator, and the visible light is detected by the semiconductor-type photoelectric converter. Pixels are square with a 143-µm pitch, which yields an image matrix of 3,136 x 3,121 (3K) pixels, with 14 bits per pixel. The design configuration is such that the detector integrates into the existing general radiographic equipment of a Bucky stand without major modification.
Image Acquisition and Display
A posteroanterior chest radiograph was obtained with the following parameters: 120 kV, 2.5 mAs, and a 12:1 antiscatter grid with a 180-cm focus-detector distance.
The digital data were sent to a PACS server (Radmax; MaroTech, Seoul, Korea) and distributed to workstations (Radmax; MaroTech, Seoul, Korea). All images were downloaded onto a local hard drive of a display workstation before interpretation. Each storage-phosphor image was 28.7 MB, and each flat-panel–detector image was 13.2 MB. A 21-inch video monitor with 2,048 x 2,560 x 8-bit pixels (model DR110; Dataray, Denver, Colo) was used in a darkened room. The monitor was operated at 71 Hz in an interlaced mode and had a maximum brightness level of 100 foot-lamberts. Because our viewing program does not support 14-bit digital images, gray-scale of the digital images, obtained with the flat-panel–detector system, was modified to 12 bits. Soft-copy images were displayed without image postprocessing such as spatial frequency enhancement. Interpreters were allowed to adjust the window width and window level of the images. Because magnification of the images was not allowed, the spatial resolution of each soft-copy image was defined by that of the monitor. The displayed image size on the monitor was the same for both detector systems.
Image Evaluation
All radiographs were analyzed by six board-certified radiologists (J.M.G., H.J.L., M.J.C., J.B.S., H.Y.K., Y.J.I.) whose levels of experience in chest radiography varied (range, 5–11 years; mean, 8.1 years). All observers were accustomed to using PACS. Training sessions were held before the scoring sessions to allow the observers to become familiar with the chest phantom and the range of simulated abnormalities, thus minimizing learning bias. A separate set of radiographs was used for the training sessions, and immediate feedback was provided during these sessions. The basic set of 40 images (20 acrylic plates x 2 detector systems) was randomized, and the images were divided into two reading subsets of 20 images each. Each observer assessed the images independently. The observers read the subsets in different orders to avoid bias. To reduce learning bias, sessions were held at least 1 week apart. No limit was imposed on reading time (mean time, 1 hour per session). The observers did not know the proportion of simulated abnormalities. A continuous rating scale of 0–100 was used to represent each observer’s confidence level regarding the presence or absence of nodules, micronodules, lines, and reticular opacities.
Data and Statistical Analysis
A total of 14,400 observations (20 acrylic plates x 15 fields x 4 lesion types x 6 observers x 2 detector systems) were evaluated. Observer performance for the detection of simulated abnormalities with the two detector systems was tested by using receiver operating characteristic (ROC) analysis of individual and averaged reader data. Detection accuracy was measured according to the area under the ROC curve, or Az, value. We used a multireader-multicase ROC approach with the jackknife method (7) to allow for generalization to the population of readers and cases (8,9). The statistical significance of the results was reported at 95% CIs for the mean differences in Az values for observer performance with use of the two detector systems (10). Mean differences were regarded as statistically significant at the 5% level when the corresponding CI did not encompass zero (10).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
Three radiologists (observers 1, 4, and 6) performed significantly better (P < .05) at detecting nodules and micronodules with the flat-panel–detector system than with the storage-phosphor system. Three radiologists (observers 2, 4, and 6) performed significantly better (P < .05) at detecting lines with the flat-panel–detector system than with the storage-phosphor system. Two radiologists (observers 2 and 6) performed significantly better (P < .05) at detecting reticular opacities with the flat-panel–detector system than with the storage-phosphor system. However, no radiologist had significantly better detection of various simulated lesions with the storage-phosphor system than with the flat-panel–detector system.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Spatial resolution in digital radiography is determined largely by the number and size of the pixels that make up an image. Because matrix size directly influences not only the spatial resolution but also the costs associated with a digital imaging system, optimum matrix size should allow sufficient spatial resolution to enable an acceptable level of diagnostic accuracy, while at the same time minimizing data acquisition, processing, transmission, and storage costs (15). In the present study, there was some difference in the pixel sizes of the two detector systems (0.1 mm for the storage-phosphor system and 0.143 mm for the flat-panel–detector system). However, because the matrix of the monitor was smaller than that of the detector systems and because magnification of soft-copy images was not allowed in our study, the spatial resolution of the two detector systems was limited by the monitor. In a ROC study by Miró et al (16), observer performance in terms of detecting parenchymal, mediastinal, and pleural abnormalities was not significantly different on 2- and 4-K storage-phosphor chest radiographs. Lams and Cocklin (17) also demonstrated that the detection of solitary pulmonary nodules by using a 1-K cathode-ray-tube monitor display did not significantly improve at effective pixel sizes of smaller than 0.8 mm. Therefore, the allowance of magnification in the current study might not have influenced observer performance for simulated lesion detection.
In the present study, the flat-panel–detector system tended to be better at depicting simulated lesions, although this was statistically significant only for nodule detection. The results of the study can be explained in several ways.
Flat-panel–detector systems permit detective quantum efficiency that exceeds the performance of storage-phosphor systems (3,5,18). Although measurements of physical imaging characteristics cannot be directly related to expected observer performance in a diagnostic setting, there is general agreement that a higher detective quantum efficiency is indicative of superior image quality, at least in terms of the fundamentals of image detection (18).
Image blurring can result from the scattering of x-ray beams, light, or both in the detector. At storage-phosphor radiography, the grain structure of the detector causes internally generated noise and a lower signal-to-noise ratio, resulting in deterioration of the image. Light is scattered in the photostimulable phosphor of the storage-phosphor system, and this produces a curved signal profile that blurs the image. Moreover, with a structured scintillator, as used in a flat-panel detector, light spreading is greatly reduced (4). Although we did not perform a systematic analysis, more prominent image noise was noted with storage-phosphor radiography (Figure).
One of the major advantages of the digital system is the wide dynamic range of the detector. These characteristics explain the improved contrast throughout the image and allow better visualization of low-contrast regions, such as the mediastinum. According to a study by Floyd et al (18), measurement of inherent contrast sensitivity showed little difference between the flat-panel–detector and storage-phosphor systems. However, because the inherent contrast of the two detectors was comparable and because the noise power spectrum of the flat-panel–detector system was far superior to that of the storage-phosphor system, one may conclude that the contrast-to-noise ratio of the former should also be superior to that of the latter (18). Chotas and Ravin (19) suggested that observer performance in a contrast-detail detection task can be improved by using images acquired with the flat-panel digital chest radiography system, as compared with those acquired with state-of-the-art screen-film combinations.
Because we evaluated soft-copy images and because adjustment of the window width and window level of the images was allowed, the difference in the gray scale of each imaging system, that is, 10 bits for the storage-phosphor system and 12 bits for the flat-panel–detector system, may also have affected observer performance.
Since one of the advantages of digital radiography is its use in adjusting image appearance for each imaging task, a limitation of our study was that we did not apply any image postprocessing. However, because there are so many image postprocessing systems, it is difficult to compare two imaging systems by using postprocessed images.
The large areas under the curves for reticular opacities suggest that the depiction of these lesions was too easy. However, the results for these lesions did not differ from those for other simulated lesions in the current study.
Finally, the clinical utility of the flat-panel–detector system can be judged on the basis of an observer study of a clinical evaluation. The results of an earlier clinical evaluation of the flat-panel–detector system, which was based on subjective judgment of the visualization of anatomic structures with soft-copy reading, indicates that the flat-panel–detector system performs better than the storage-phosphor system (6).
In conclusion, this phantom study with use of simulated lesions revealed that the amorphous silicon detector system appears to be superior to the storage-phosphor system for the detection of pulmonary nodules at soft-copy image evaluation.
Practical application: In the evaluation of soft-copy images, the superior performance with use of the flat-panel–detector system in the detection of simulated pulmonary nodules and the generally equivalent performance in the detection of other simulated abnormalities suggest the potential for improved patient care.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, J.M.G., J.G.I., J.H.K.; study concepts and design, J.M.G., J.G.I., J.H.K.; literature research, J.M.G., H.J.L., M.J.C.; experimental studies, J.M.G., H.J.L., J.W.K.; data acquisition, J.M.G., H.J.L., J.W.K.; data analysis/interpretation, J.M.G., J.G.I., H.J.L., M.J.C., J.B.S., H.Y.K., Y.J.L.; statistical analysis, J.M.G., J.H.K.; manuscript preparation, J.M.G.; manuscript definition of intellectual content, J.M.G., J.G.I., J.H.K.; manuscript editing, J.G.I., J.H.K.; manuscript revision/review and final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Korner, C. H. Weber, S. Wirth, K.-J. Pfeifer, M. F. Reiser, and M. Treitl Advances in Digital Radiography: Physical Principles and System Overview RadioGraphics, May 1, 2007; 27(3): 675 - 686. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Don, B. R. Whiting, J. S. Ellinwood, D. H. Foos, K. A. Kronemer, and R. A. Kraus Neonatal Chest Computed Radiography: Image Processing and Optimal Image Display Am. J. Roentgenol., April 1, 2007; 188(4): 1138 - 1144. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Bacher, P. Smeets, L. Vereecken, A. De Hauwere, P. Duyck, R. De Man, K. Verstraete, and H. Thierens Image quality and radiation dose on digital chest imaging: comparison of amorphous silicon and amorphous selenium flat-panel systems. Am. J. Roentgenol., September 1, 2006; 187(3): 630 - 637. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. W. Hamer, C. B. Sirlin, M. Strotzer, I. Borisch, N. Zorger, S. Feuerbach, and M. Volk Chest Radiography with a Flat-Panel Detector: Image Quality with Dose Reduction after Copper Filtration Radiology, November 1, 2005; 237(2): 691 - 700. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K Ono, T Yoshitake, K Akahane, Y Yamada, T Maeda, M Kai, and T Kusama Comparison of a digital flat-panel versus screen-film, photofluorography and storage-phosphor systems by detection of simulated lung adenocarcinoma lesions using hard copy images Br. J. Radiol., October 1, 2005; 78(934): 922 - 927. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. J. M. Kroft, W. J. H. Veldkamp, B. J. A. Mertens, M. V. Boot, and J. Geleijns Comparison of Eight Different Digital Chest Radiography Systems: Variation in Detection of Simulated Chest Disease Am. J. Roentgenol., August 1, 2005; 185(2): 339 - 346. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Goo, J.-Y. Choi, J.-G. Im, H. J. Lee, M. J. Chung, D. Han, S. H. Park, J. H. Kim, and S.-H. Nam Effect of Monitor Luminance and Ambient Light on Observer Performance in Soft-Copy Reading of Digital Chest Radiographs Radiology, September 1, 2004; 232(3): 762 - 766. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
